Neurology · 2025
Перинатальное психическое заболевание у матерей с множественным склерозом и другими хроническими заболеваниями в Онтарио, Канада.
СтатьяНазвание и аннотация: Машинный перевод · не проверен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Marrie Ruth Ann, Bolton James, Ling Yushu Vicki, Bernstein Charles, Krysko Kristen M, Li Ping, Mckay Kyla A, Pequeno Priscila, Razaz Neda, Rotstein Dalia, Deakin-Harb Karma, Maxwell Colleen J
Аннотация
СОСТОЯНИЕ И ЦЕЛИ: Перинатальные расстройства настроения и тревоги являются наиболее частой формой материнской заболеваемости среди населения в целом, однако мало что известно о перинатальных психических заболеваниях у матерей с множественным склерозом (МС). Мы сравнили частоту и распространенность перинатальных психических заболеваний у матерей с склерозом, эпилепсией, воспалительным кишечным заболеванием (ИБД), диабетом и женщин без этих заболеваний.МЕТОДЫ: Используя взаимосвязанные данные об административном здоровье населения, полученные из ON, Канада, мы провели когортное исследование матерей, имеющих MS, эпилепсию, IBD и диабет, и без этих заболеваний (сопоставителей), которые имели живое рождение с индексными датами, определяемыми как 1 год до зачатия, между 2002 и 2017. С использованием подтвержденных определений мы оценивали распространенность и распространенность психических заболеваний (каких-либо, депрессия, тревога, биполярное расстройство, психоз, употребление наркотиков, попытка самоубийства) в дородовой период (от зачатия до рождения) и 3 годы после родов. Мы сравнили оценки частотности и распространенности среди групп населения с использованием простых коэффициентов заболеваемости (ОР) и коэффициентов распространенности с 95% CI и использование моделей регрессии Poisson, корректирующихся на смешения. Результаты: Мы включили 894,852 матери1,745 с НР; 5,954 с эпилепсией; 4,924 с IBD; 13,002 с диабетом; 869,227 Компараторы). На момент зачатия средний возраст матери (SD) составлял 28.6 (В тыс.5.7) годы. Психическое заболевание, вызванное любым инцидентом 8.4% Матери с ПМР в дородовой и послеродовой периоды 14.2% В течение первого послеродового года наиболее распространенными расстройствами были депрессия и беспокойство. Первый послеродовый год был более рискованным, чем период PN (любые психические заболевания IR) 1.27; 95% CI 1.08-1.50). После корректировки среди матерей, имеющих MS, наблюдалось увеличение числа психических заболеваний во время НР (IR) 1.26; 95% CI 1.11-1.44) и послеродовой период (IR) 1.33; 95% CI 1.20-1.47, первый послеродовой год) периоды, чем матери-компараторы. Аналогичным образом, среди матерей, имеющих НР, возросло число случаев всех конкретных психических заболеваний, за исключением попыток самоубийства в период ПН по отношению к матерям-компараторам. Любые распространенные психические заболевания 42% Матери с ПМР в дородовой и послеродовой периоды 50.3% в первый послеродовой год. Дискуссия: Матери, имеющие НР, имеют повышенную частоту и распространенность перинатальных психических заболеваний по сравнению с матерями-компараторами, хотя остаточные смешения не могут быть исключены. В этих выводах подчеркивается необходимость принятия профилактических мер и раннего лечения психических заболеваний.
Оригинал публикации
Peripartum Mental Illness in Mothers With Multiple Sclerosis and Other Chronic Diseases in Ontario, Canada.
BACKGROUND AND OBJECTIVES: Peripartum mood and anxiety disorders constitute the most frequent form of maternal morbidity in the general population, but little is known about peripartum mental illness in mothers with multiple sclerosis (MS). We compared the incidence and prevalence of peripartum mental illness among mothers with MS, epilepsy, inflammatory bowel disease (IBD), and diabetes and women without these conditions.
METHODS: Using linked population-based administrative health data from ON, Canada, we conducted a cohort study of mothers with MS, epilepsy, IBD, and diabetes and without these diseases (comparators) who had a live birth with index dates, defined as 1 year before conception, between 2002 and 2017. Using validated definitions, we estimated the incidence and prevalence of mental illness (any, depression, anxiety, bipolar disorder, psychosis, substance use, suicide attempt) during the prenatal (PN) period (from conception to birth) and 3 years postpartum. We compared incidence and prevalence estimates between cohorts using simple incidence ratios (IRs) and prevalence ratios with 95% CIs and using Poisson regression models adjusting for confounders.
RESULTS: We included 894,852 mothers (1,745 with MS; 5,954 with epilepsy; 4,924 with IBD; 13,002 with diabetes; 869,227 comparators). At conception, the mean (SD) maternal age was 28.6 (5.7) years. Any incident mental illness affected 8.4% of mothers with MS prenatally and 14.2% during the first postpartum year; depression and anxiety were the most common incident disorders. The first postpartum year was a higher risk period than the PN period (any mental illness IR 1.27; 95% CI 1.08-1.50). After adjustment, mothers with MS had an increased incidence of any mental illness during the PN (IR 1.26; 95% CI 1.11-1.44) and postpartum (IR 1.33; 95% CI 1.20-1.47, first postpartum year) periods than comparator mothers. Similarly, mothers with MS had an increased incidence of all specific mental illnesses except suicide attempt during the PN period vs comparator mothers. Any prevalent mental illness affected 42% of mothers with MS prenatally and 50.3% in the first postpartum year.
DISCUSSION: Mothers with MS had an elevated incidence and prevalence of peripartum mental illness compared with comparator mothers, although residual confounding cannot be excluded. These findings emphasize the need for preventive interventions and early treatment of mental illness.
Перевод полного текста: Перевод в очереди
Читать найденный полный текст
This is an open access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND) , where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.
Introduction
Multiple sclerosis (MS) is a chronic immune-mediated inflammatory disease of the CNS, affecting nearly 3 million individuals worldwide, of whom three-quarters are women. 1 Presentation typically occurs during a period when many individuals are making reproductive decisions. Women with MS have a markedly increased incidence and prevalence of depression, anxiety disorders, and bipolar disorder compared with women without MS. 2 , 3 Peripartum mood disorders constitute the most frequent form of maternal morbidity in the general population 4 and pose health risks for those affected and their families 5 ; however, little is known about the risk of peripartum mental illness in women with MS. A 2017 systematic review of the association between maternal chronic medical conditions and peripartum mental illness identified only 16 studies conducted from 1988 to 2017 6 ; only 1 study included women with MS (n = 255) and focused solely on combined mood and anxiety disorders. 7 A recent scoping review regarding women's health in MS 8 identified only 1 other small study (in Norway) that has directly addressed peripartum mental illness in MS. 9 Among 140 women diagnosed with MS before pregnancy, 15% had depression in their third trimester (based on a screening tool), two-fold more than women without MS; the prevalence of anxiety did not differ. 9 The rising prevalence of MS 1 and number of persons with MS having children 10 and potential adverse effects of peripartum mental illness on the parent and child 5 , 11 make it critical to understand the incidence and prevalence of a broad spectrum of peripartum mental illness in this population.
We compared the incidence and prevalence of peripartum mental illness in women with MS, epilepsy, inflammatory bowel disease (IBD), and diabetes and women without any of these conditions. We chose these conditions because they allow for comparisons on nondemographic factors that may affect the risk of peripartum mental illness including organ system (CNS [MS, epilepsy] vs non-CNS [IBD, diabetes]), pathophysiology (immune-mediated [MS, IBD] vs not [epilepsy, diabetes]), and disease course (relapsing remitting with underlying progression [MS and IBD] vs progressive [diabetes]). We hypothesized that incidence and prevalence of peripartum mental illness among women with MS would exceed those among women in all comparator cohorts except epilepsy, another CNS disorder, and that the incidence of mental illness would be higher in the first postpartum year than in the prenatal (PN) period. 12
Methods
Setting
We conducted this retrospective cohort study in the province of Ontario, which is located in east-central Canada and is home to nearly 40% of the Canadian population (>14 million residents). 13 Ontario has a universal, publicly funded health system with longitudinal, prospectively collected, high-quality records of health service use. 14
Data Sources
Administrative databases capture health service use and are held at ICES (formerly known as the Institute for Clinical and Evaluative Sciences). The Registered Persons Database (RPDB) provided a unique identifier, birth date, sex, region of residence (postal code), and health insurance coverage dates. Postal codes can be linked to census data to determine household income of the individual's enumeration area. The Discharge Abstract Database (DAD) provided hospital admission and separation dates and discharge diagnoses listed using 5-digit International Classification of Diseases (ICD)-9 or 10-Canadian Modification (CA) codes depending on the year. The Ontario Health Insurance Plan (OHIP) database provided dates of physician claims coded using a three-digit ICD-9 code for 1 physician-assigned diagnosis. Physicians paid through alternate funding sources submit shadow billings. 15 The Ontario Mental Health Reporting System captures admissions (discharge diagnoses) to designated adult inpatient mental health beds in general/specialty facilities. 16 Prescription claims data were not accessed because they are not population-based and do not provide indications for use.
The MOMBABY data set has captured all hospital births in Ontario since 1988. As of 2002, mothers and their newborns were deterministically linked by the DAD using the medical chart number. 17 This data set provided gestational age at delivery and indicated whether the birth was live/stillbirth and singleton/multiple. All data sets were linked at the individual level using unique encoded identifiers and analyzed at ICES.
Standard Protocol Approvals, Registrations, and Patient Consents
We obtained ethics approval from the University of Manitoba Health Research Ethics Board. Data access (through ICES) was authorized under section 45 of ON's Personal Health Information Protection Act.
Study Populations
We identified separate chronic disease cohorts (MS, epilepsy, IBD, and diabetes) using validated administrative case definitions 18 – 21 (eTable 1) using all available hospital and OHIP claims for the period April 01, 1988, and December 31, 2020, and excluded anyone missing date of birth or sex. For each chronic disease algorithm, we designated the earliest ICD-9 or 10-CA code identified as the diagnosis date. If women had more than one chronic disease, they were assigned to the disease with the earliest diagnosis date. We linked these cohorts to the MOMBABY data set to identify mothers with a delivery date during the period of interest. Mothers in the MOMBABY data set with a delivery date in the period of interest who did not have MS, epilepsy, IBD, or diabetes were designated as comparators. For each mother, we selected the first pregnancy in the study period. We estimated the conception date by subtracting clinically assessed gestational age, based on a combination of ultrasound and last menstrual period, recorded at the delivery hospitalization from the newborn's birthdate; <0.3% of all records are missing gestational age because this is a mandatory field. 22 The index date was 365 days before the estimated date of conception for the first pregnancy observed ( Figure 1 for design). We further excluded any woman (1) who had an index date before January 1, 2000, or after March 31, 2017; (2) who did not have a valid registration number; (3) whose newborn did not have a valid registration number; (4) who was not a resident in the province and eligible for provincial health insurance for ≥12 months before the index date to allow complete one-year lookback period; (5) with missing sex; (6) younger than 15 years or older than 49 years at conception date; (7) who died before the delivery date; or (7) who had a birth at gestational age <22 weeks (eFigure 1).
Study Design and Periods
Study Period
The earliest index date was January 1, 2000, and the latest was March 31, 2017. This allowed a 1-year lookback period to define covariates (income, rural residence, annual physician visits) at the index date and a 3-year (1,097 days) follow-up period after the delivery date. For each woman, we defined several intervals: the year before conception (prepregnancy), pregnancy (PN period), first year after delivery (postnatal/postpartum), and second year and third year after delivery ( Figure 1 ). Although previous studies of peripartum mental illness in the general population have reported prevalence and, to a lesser extent, incidence estimates, 12 , 23 few have reported which part of the peripartum interval carries the greatest risk. Therefore, we also divided the PN period into trimesters where trimester 1 was from conception date to 13 weeks 6 days, trimester 2 was from 14 weeks to 26 weeks 6 days, and trimester 3 was from 27 weeks to delivery.
Peripartum Mental Illness
We applied validated case definitions for identifying depression, anxiety, bipolar disorder, psychosis, substance use disorder, suicide attempt, and any mental illness (encompassing any of the listed mental illnesses or mental health service use) (eTable 2). The any mental illness definition has a positive predictive value (PPV) of 84.9% and negative predictive value of 96% (eTable 3). This sensitive definition relies on a single diagnosis or mental illness service code to identify mental illness, consistent with the approach of the Canadian Chronic Disease Surveillance System. 24 We used a similar approach for the other mental illnesses; most case definitions exceeded the recommended PPV of 70% (eTable 3). 25 , 26 We also used a specific definition that required multiple diagnosis codes over a specified period (eTable 2).
If a mother met a case definition, we assumed once affected, always affected, in keeping with relapsing-remitting nature of these conditions. To identify mental illness as incident for the PN period, there had to be no diagnosis codes for the mental illness of interest for 2 years before the conception date. For incident mental illness in the postnatal period, there had to be no diagnosis codes for the mental illness of interest for 2 years before the conception date and during the PN period. The 2-year window was based on our previous observation that in immune-mediated inflammatory disease cohorts, among those who met the case definition for depression, the median [interquartile range (IQR)] time between depression claims was 24.0 (36.0–121.0) days; the 99th percentile of the distribution for time between claims was 497 days. 27
Characteristics
Sociodemographic characteristics assessed at index date included income quintile based on postal code and urban vs rural residence based on postal code, where census metropolitan areas with populations ≥100,000 were designated as urban. Maternal age was assessed at conception. Information about ethnicity was not available in the data sets. Pregnancy characteristics included delivery year, parity (any pregnancies before the study period), multiple births, and maternal comorbidity. Comorbidity was assessed using the Obstetric Comorbidity Index, 28 a weighted index. We modified the index to exclude gestational diabetes, to avoid misclassification in the diabetes cohort, and alcohol and drug abuse, because substance use was an outcome. We summarized health care use as the annual number of physician visits at the index and conception dates. We also assessed access to PN care by (1) use of PN ultrasound (present/absent) and (2) number of PN visits for obstetrics reasons, whether to primary care providers or obstetricians.
Analysis
We summarized patient characteristics using means (SDs), median (IQR), and frequency (percentage). We compared characteristics between groups using χ 2 tests, Kruskal-Wallis tests, and analysis of variance (ANOVA). Owing to the large sample size, which may lead to the identification of statistically significant differences between groups that are not clinically meaningful, we show standardized differences rather than p values. Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large. 29
We estimated the crude and age-standardized incidence proportion and prevalence of mental illness during the PN period overall and for each trimester and in the first, second, and third postpartum years. Estimates were standardized to the 2015 Canadian population using the direct method. We compared the age-standardized estimates between cohorts using incidence ratios and prevalence ratios (PRs) and 95% CIs based on bootstrapping with 500 repetitions.
We used a multivariable Poisson regression model to compare the incidence rates between the PN and postnatal periods and between disease groups, using the natural log of person-time as the model offset and scaling the deviance. Covariates included age at conception (15–24, 25–34, ≥35 [reference]), delivery year (2002–2009 [reference], 2010–2019), income (3 lowest quintiles vs 2 highest because some outcomes were rare), and Obstetric Comorbidity Index (0, 1, 2, ≥3). We similarly modeled prevalence using generalized estimating equations with an exchangeable correlation matrix to account for repeated observations over the study period. We limited these regression analyses to any mental illness, depression, and anxiety, given the small cell sizes for other outcomes. We report rate ratios (RRs) and 95% CI.
Complementary Analysis
Given potential differences in the types of health care providers seen and potentially differential recognition of mental illness, we summarized the type of specialist who submitted the first diagnostic claim for the illness (e.g., depression) as follows: primary care provider, psychiatrist, and other specialists. Individuals who were not diagnosed in outpatient settings were categorized as non-OHIP. E-value analysis was used as a sensitivity analysis to show the strength of association required with the disease group and the outcome for an unmeasured confounder to fully explain the observed association and is reported in footnotes of relevant tables and figures.
Statistical analyses used SAS V9.4 (SAS Institute Inc., Cary, NC).
Data Availability
The data set from this study is held securely in coded form at ICES. Legal data-sharing agreements prohibit ICES from making the data set publicly available. However, access may be granted to those who meet prespecified criteria for confidential access, available at ices.on.ca/DAS (email das@ices.on.ca ). The full data set creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are, therefore, either inaccessible or may require modification.
Results
Cohorts
We included 894,852 mothers with live births, 1,745 with MS, 5,954 with epilepsy, 4,924 with IBD, and 13,002 with diabetes; the remaining 869,227 were comparator mothers ( Table 1 , eFigure 1). Among the disease cohort mothers, only 2.9% had more than one of the health conditions of interest. Among mothers with MS, 14 had IBD, 36 had epilepsy, and 35 had diabetes. At conception, mothers with MS were similar in age to mothers with diabetes (average difference = 0.03 years) and slightly older than those with IBD (0.91 years) but substantially older than mothers with epilepsy (3.79 years) and comparator mothers (2.62 years). Mothers with diabetes had the highest comorbidity index scores.
Table 1 Sociodemographic Characteristics of Participants Characteristic Total MS Epilepsy IBD Diabetes Comparators Standardized difference (comparator group as reference) MS Epilepsy IBD Diabetes No. of women 894,852 1,745 5,954 4,924 13,002 869,227 No. of pregnancies during the study period, n (%) 1 422,341 (47.2) 970 (55.6) 3,013 (50.6) 2,393 (48.6) 7,862 (60.5) 408,103 (47.0) 0.17 0.07 0.03 0.27 2 364,651 (40.7) 661 (37.9) 2,248 (37.8) 2,067 (42.0) 4,237 (32.6) 355,438 (40.9) 0.06 0.06 0.03 0.17 ≥3 86,972 (9.7) 114 (6.5) 693 (11.7) 464 (9.4) 903 (6.9) 105,686 (12.2) 0.20 0.02 0.09 0.18 Any pregnancy before index date, n (%) 53,549 (6.0) 119 (6.8) 415 (7.0) 298 (6.1) 2,117 (16.3) 50,510 (5.8) 0.03 0.04 0 0.34 Multiple births (yes) for first pregnancy, n (%) 17,167 (1.9) 51 (2.9) 126 (2.1) 121 (2.5) 300 (2.3) 16,569 (1.9) 0.07 0.01 0.04 0.03 Multiple births (yes) for any pregnancies, n (%) 26,777 (3.0) 73 (4.2) 187 (3.1) 179 (3.6) 409 (3.1) 25,929 (3.0) 0.06 0.01 0.04 0.01 Delivery year (calendar), n (%) 2002–2005 153,932 (17.2) 297 (17.0) 842 (14.1) 770 (15.6) 2,428 (18.7) 149,595 (17.2) 0 0.08 0.04 0.04 2006–2009 244,475 (27.3) 469 (26.9) 1,496 (25.1) 1,266 (25.7) 3,621 (27.8) 237,623 (27.3) 0.01 0.05 0.04 0.01 2010–2013 226,884 (25.4) 462 (26.5) 1,594 (26.8) 1,295 (26.3) 3,202 (24.6) 220,331 (25.3) 0.03 0.03 0.02 0.02 2014–2019 269,561 (30.1) 517 (29.6) 2022 (34.0) 1,593 (32.4) 3,751 (28.8) 261,678 (30.1) 0.01 0.08 0.05 0.03 Age at conception date (yrs), mean (SD) 28.63 (5.7) 31.21 (4.8) 27.42 (5.9) 30.30 (4.8) 31.24 (5.8) 28.59 (5.7) 0.49 0.21 0.31 0.45 Age group at conception, n (%) 15–19 64,468 (7.2) 11 (0.6) 626 (10.5) 61 (1.2) 387 (3.0) 63,383 (7.3) 0.34 0.12 0.3 0.2 20–24 143,209 (16.0) 126 (7.2) 1,319 (22.2) 467 (9.5) 1,228 (9.4) 140,069 (16.1) 0.28 0.16 0.2 0.2 25–29 285,509 (31.9) 509 (29.2) 1,827 (30.7) 1,634 (33.2) 3,269 (25.1) 278,270 (32.0) 0.06 0.03 0.03 0.15 30–34 265,135 (29.6) 652 (37.4) 1,457 (24.5) 1,824 (37.0) 4,262 (32.8) 256,940 (29.6) 0.16 0.12 0.16 0.07 35–39 112,066 (12.5) 377 (21.6) 598 (10.0) 779 (15.8) 2,930 (22.5) 107,382 (12.4) 0.24 0.08 0.1 0.27 40–49 a 24,465 (2.8) 70 (4.0) 127 (2.1) 159 (3.2) 926 (7.1) 23,183 (2.7) 0.07 0.04 0.20 0.03 Region at index, n (%) Urban 807,467 (90.2) 1,601 (91.7) 5,246 (88.1) 4,377 (88.9) 11,863 (91.2) 784,380 (90.2) 0.05 0.07 0.04 0.03 Rural 87,385 (9.8) 144 (8.3) 708 (11.9) 537 (10.9) 1,149 (8.8) 84,847 (9.8) 0.05 0.07 0.04 0.03 Income quintile at index, n (%) Quintile 1 (lowest) 201,140 (22.5) 288 (16.5) 1,550 (26.0) 760 (15.4) 3,253 (25.0) 195,289 (22.5) 0.15 0.08 0.18 0.06 Quintile 2 189,009 (21.1) 344 (19.7) 1,234 (20.7) 909 (18.5) 2,776 (21.4) 183,746 (21.1) 0.04 0.01 0.07 0.01 Quintile 3 185,826 (20.8) 388 (22.2) 1,225 (20.6) 1,071 (21.8) 2,820 (21.7) 180,322 (20.7) 0.04 0 0.02 0.02 Quintile 4 175,579 (19.6) 380 (21.8) 1,094 (18.4) 1,136 (23.1) 2,407 (18.5) 170,562 (19.6) 0.05 0.03 0.08 0.03 Quintile 5 143,298 (16.0) 297 (17.0) 851 (14.3) 1,048 (21.3) 1,746 (13.4) 139,308 (16.0) 0.1 0.05 0.14 0.07 Physician visits, median (IQR) 15 (12–18) 16 (12–19) 16 (13–20) 16 (13–19) 17 (14–21) 15 (12–18) 0.08 0.15 0.13 0.34 Prenatal visits, median (IQR) 15 (12–18) 16 (12–19) 16 (13–20) 16 (13–19) 17 (14–21) 15 (12–18) 0.08 0.15 0.13 0.34 Any prenatal ultrasound, n (%) 785,651 (87.8) 1,586 (90.9) 5,410 (90.0) 4,512 (91.6) 12,092 (93.0) 762,051 (87.7) 0.1 0.1 0.13 0.18 Obstetric Comorbidity Index score at conception, mean (SD) 0.05 (0.31) 0.05 (0.31) 0.08 (0.38) 0.07 (0.37) 0.12 (0.47) 0.05 (0.31) 0 0.07 0.06 0.17 Obstetric Comorbidity Index score at conception, n (%) a 0 858,243 (95.9) 1,678 (96.2) 5,614 (94.3) 4,663 (94.7) 11,861 (91.2) 834,427 (96.0) 0.01 0.08 0.06 0.2 1 30,493 (3.4) 52 (3.0) 274 (4.6) 205 (4.2) 929 (7.1) 29,033 (3.3) 0.02 0.06 0.04 0.17 ≥2 6,116 (0.6) 15 (0.8) 66 (1.2) 56 (1.2) 212 (1.7) 5,767 (0.6) 0.02 0.06 0.06 0.10 Abbreviations: IBD = inflammatory bowel disease; IQR = interquartile range; MS = multiple sclerosis. Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large. a Upper categories collapsed to avoid small cells of <6 and preserve privacy.
Sociodemographic Characteristics of Participants
Abbreviations: IBD = inflammatory bowel disease; IQR = interquartile range; MS = multiple sclerosis.
Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large.
Upper categories collapsed to avoid small cells of <6 and preserve privacy.
Incident Mental Illness in the MS Cohort
Because age-standardized findings were similar to those from multivariable regression, we present the regression results throughout. The highest crude incidence proportion was for any mental illness; 8.4% of mothers were affected prenatally and 14.2% in the first postpartum year. Among individual disorders, focusing on the specific (more conservative) definitions, the condition with the highest incidence during the PN period was anxiety (3.8%), followed by depression (0.82%) and substance use (0.54%); estimates were higher for sensitive definitions ( Figure 2 ), and no suicide attempts were recorded. Similarly, the pattern of which conditions had the highest and lowest incidence was generally the same during the first, second, and third postpartum years as during the PN period, with the exception of psychosis, where cell sizes were small, and bipolar disorder, which was more common than substance use postpartum. We observed a general pattern of incident mental illness in which the highest risk period was the first postpartum year for all conditions, followed by the PN period; this persisted after multivariable adjustment ( Figure 2 , eTable 4).
Crude Incidence Proportion of Mental Illness by Pregnancy Period in Mothers With Multiple Sclerosis
*Where 2 definitions were tested, only the sensitive definition is presented here. Empty cells are shown where 6 or fewer events were observed to preserve privacy and confidentiality.
Incident Mental Illness in the MS Cohort vs Comparators
In the comparator cohort, the incidence proportion of any mental illness was 7.06% during the PN period and the incidence of anxiety was 2.28% and of depression was 0.95%. Compared with mothers in the comparator cohort, mothers with MS had a higher incidence of PN mental illness (any), and for all individual conditions except suicide attempt for which the sample was too small to estimate accurately ( Figure 3 ). The highest relative ratios were observed for the psychosis definition, followed by the specific definitions of depression and bipolar disorder ( Figure 3 ); ratios were generally higher for specific than sensitive definitions. Incidence proportions among mothers with MS remained higher than in comparator mothers in the first, second, and third postpartum years except for any mental illness in the third postpartum year ( Figure 4 ).
Association Between Mothers With MS, Epilepsy, IBD, and Diabetes, Compared With Controls and Incident Mental Illness in the Prenatal Period (Adjusted Rate Ratios, 95% CI)* , ^
*Adjusted for age at conception, delivery year, income quintile, and Obstetric Comorbidity Index. Where sensitive and specific definitions were tested, only specific definitions were shown for visual simplicity. No suicide attempts were recorded for multiple sclerosis. ^E-value analysis for the smallest statistically significant association of RR 1.13 (lower confidence limit): 1.51 (1.49). Rate ratios of 1.25 have an E-value of 1.81, rate ratios of 1.50 have an E-value of 2.37, and rate ratios of 2.0 have an E-value of 3.41. IBD = inflammatory bowel disease; MS = multiple sclerosis.
Association Between MS and Incident Mental Illness During Prenatal (PN) and First 3 Postpartum Years (PYs), Compared With Controls (Adjusted Rate Ratios, 95% CI)* , ^
*Adjusted for age at conception, delivery year, income quintile, and Obstetric Comorbidity Index. The strength of association is generally higher for specific than sensitive definitions. Suicide attempt was suppressed because of some cells <6. ^E-value analysis for the smallest statistically significant association of RR 1.26 (lower confidence limit): 1.83 (1.46). Rate ratios of 1.50 have an E-value of 2.37, and rate ratios of 2.0 have an E-value of 3.41. IBD = inflammatory bowel disease; MS = multiple sclerosis; Sens = sensitive; Spec = specific.
Other Disease Cohorts
During the PN period, the crude incidence proportion of mental illness and multivariable-adjusted incidence rates were increased in all other disease cohorts for all mental illnesses evaluated ( Figure 3 for adjusted RRs prenatally, eFigure 2 for adjusted RRs in the first postpartum year), with 2 exceptions, namely the incidence proportion of psychosis and suicide attempts in the IBD cohort.
Comparisons Between MS and Other Disease Cohorts: Incident Mental Illness
During the PN period, the MS cohort had an increased incidence of depression (sensitive definition), anxiety, psychosis (specific definition), and substance use compared with the IBD and diabetes cohorts ( Table 2 , eTable 5). The MS cohort also had an increased incidence of depression (both definitions, although not statistically significant for the specific definition) and psychosis (specific definition) compared with the epilepsy cohort during the PN period. Postpartum, the MS cohort continued to demonstrate a general pattern of higher incidence of depression, anxiety, and psychosis than mothers with IBD or diabetes. Findings were more variable when compared with the epilepsy cohort, being similar (any mental illness), higher (depression specific definition, postpartum year 2, bipolar disorder, postpartum year 3), and lower (bipolar disorder specific definition and substance use, postpartum years 1 and 2; psychosis, postpartum years 1–3).
Table 2 Comparison of Incidence Rates of Mental Illness Between Multiple Sclerosis (MS) and Other Disease Cohorts a , b Disease cohort Comparator cohort Prenatal Postpartum y 1 Postpartum y 2 Postpartum y 3 Any mental illness Multiple sclerosis Epilepsy 0.93 (0.81, 1.08) 1.05 (0.93, 1.18) 1.17 (1.00, 1.37) 0.97 (0.81, 1.16) Multiple sclerosis Inflammatory bowel disease 1.12 (0.96, 1.30) 1.03 (0.91, 1.16) 1.08 (0.93, 1.27) 0.85 (0.70, 1.02) Multiple sclerosis Diabetes 1.07 (0.93, 1.23) 1.16 (1.04, 1.29) 1.20 (1.04, 1.39) 1.03 (0.86, 1.22) Depression (specific) Multiple sclerosis Epilepsy 1.15 (0.99, 1.32) 0.98 (0.89, 1.08) 1.20 (1.05, 1.38) 1.14 (0.97, 1.33) Multiple sclerosis Inflammatory bowel disease 1.60 (1.36, 1.87) 1.65 (1.48, 1.84) 1.59 (1.37, 1.85) 0.99 (0.85, 1.16) Multiple sclerosis Diabetes 1.14 (0.99, 1.31) 1.42 (1.29, 1.56) 1.10 (0.97, 1.33) 1.53 (1.32, 1.78) Anxiety Multiple sclerosis Epilepsy 1.07 (0.94, 1.21) 1.24 (1.11, 1.38) 1.03 (0.89, 1.19) 1.07 (0.89, 1.29) Multiple sclerosis Inflammatory bowel disease 1.25 (1.09, 1.43) 0.98 (0.88, 1.09) 1.18 (1.01, 1.37) 1.25 (1.03, 1.52) Multiple sclerosis Diabetes 1.25 (1.10, 1.41) 1.20 (1.08, 1.33) 1.11 (0.97, 1.28) 1.35 (1.13, 1.60) Bipolar disorder (specific) Multiple sclerosis Epilepsy 0.82 (0.71, 0.95) 0.63 (0.55, 0.72) 0.43 (0.34, 0.50) 2.21 (1.90, 2.57) Multiple sclerosis Inflammatory bowel disease 1.20 (1.03, 1.41) 1.02 (0.89, 1.18) 0.48 (0.38, 0.61) 2.17 (1.85, 2.55) Multiple sclerosis Diabetes 1.09 (0.94, 1.26) 1.07 (0.94, 1.22) 0.40 (0.32, 0.50) 1.66 (1.45, 1.90) Psychosis (specific) Multiple sclerosis Epilepsy 1.32 (1.14, 1.54) 1.03 (0.92, 1.14) 0 0.33 (0.27, 0.40) Multiple sclerosis Inflammatory bowel disease 5.76 (4.58, 7.25) 2.89 (2.53, 3.30) 0 0.58 (0.47, 0.71) Multiple sclerosis Diabetes 1.47 (1.27, 1.70) 2.08 (1.87, 2.31) 0 1.19 (0.97, 1.45) Substance use Multiple sclerosis Epilepsy 1.08 (0.93, 1.26) 0.56 (0.46, 0.68) 0.73 (0.61, 0.88) 1.14 (0.99, 1.31) Multiple sclerosis Inflammatory bowel disease 1.81 (1.51, 2.16) 0.60 (0.49, 0.74) 0.84 (0.69, 1.03) 1.61 (1.38, 1.89) Multiple sclerosis Diabetes 1.25 (1.08, 1.46) 0.52 (0.43, 0.63) 0.88 (0.74, 1.06) 1.32 (1.15, 1.51) Abbreviation: RR = rate ratio. a Adjusted for age group, socioeconomic status, delivery year, and Obstetric Comorbidity Index. Rate ratio above 1 indicates increased incidence of the mental illness of interest during the peripartum interval of interest (e.g., prenatal) in mothers with MS vs the comparator disease. Comparisons with controls are given in Figure 4 . b E-value analysis for the smallest statistically significant association of RR 1.16 (lower confidence limit): 1.59 (1.24). Estimates of 1.25 have an E-value of 1.81, of 1.50 have an E-value of 2.37, and of 2.0 have an E-value of 3.41.
Comparison of Incidence Rates of Mental Illness Between Multiple Sclerosis (MS) and Other Disease Cohorts a , b
Abbreviation: RR = rate ratio.
Adjusted for age group, socioeconomic status, delivery year, and Obstetric Comorbidity Index.
Rate ratio above 1 indicates increased incidence of the mental illness of interest during the peripartum interval of interest (e.g., prenatal) in mothers with MS vs the comparator disease. Comparisons with controls are given in Figure 4 .
E-value analysis for the smallest statistically significant association of RR 1.16 (lower confidence limit): 1.59 (1.24). Estimates of 1.25 have an E-value of 1.81, of 1.50 have an E-value of 2.37, and of 2.0 have an E-value of 3.41.
Prevalent Mental Illness in the MS Cohort Compared With Other Cohorts
When we examined prevalence estimates, 42% of mothers with MS had any mental illness during the PN period ( Figure 5 for crude estimates). In the first postpartum year, the prevalence of any mental illness was 50.3%; the most prevalent specific conditions were anxiety (25.1%), followed by depression (4.6%). Notably, approximately 1% of mothers had psychosis, and nearly 6% had substance use disorders in postpartum year 1. Compared with mothers in the comparator cohort, the adjusted prevalence of PN mental illness was generally higher among mothers with MS (eFigure 3), consistent with the findings for incidence. The highest PRs were for substance use and bipolar disorder, followed by psychosis. The prevalence of suicide attempt was reduced. The prevalence of mental illness in the first postpartum year was higher in all non-MS disease cohorts compared with the comparator cohort for all mental illnesses studied with 1 exception (eFigure 4). The prevalence of psychosis (sensitive and specific definitions) was higher in all cohorts except in the IBD cohort, in which it was reduced; however, this should be interpreted with caution because of small numbers.
Crude Period Prevalence of Mental Illness by Pregnancy Period in Mothers With Multiple Sclerosis
Where 2 definitions were tested, the sensitive definition is presented for simplicity. Y= year.
Complementary Analysis
Initial mental illness diagnoses were mostly recorded by family physicians and psychiatrists; although these proportions varied by mental illness, distributions were similar across disease cohorts (eFigure 5).
Discussion
In this large population-based cohort study, we investigated the incidence and prevalence of the spectrum of peripartum mental illness in Ontario women with MS, epilepsy (another CNS disease), IBD (another immune-mediated disease), and diabetes relative to comparators without these conditions. The incidence was nonspecifically increased during the PN and postpartum periods among all disease cohorts examined, affecting nearly all specific mental illnesses examined. Mothers with MS had an increased incidence of depression, anxiety, bipolar disorder, psychosis, and substance use than mothers with IBD and diabetes during the PN period. As hypothesized, the incidence of mental illness was higher in the first postpartum year than during the PN period and continued to decline in the second and third postpartum years.
In a 2017 systematic review, the estimated prevalence of PN depression in high-income countries was 9.2% (8.4%–10%) and of postnatal depression was 9.5% (8.9%–10.1%), 4 comparable to the findings in our comparator cohort. The incidence of PN depression was 13%, higher than we observed, and the estimates of the incidence of postnatal depression ranged widely from 5% to 25%. 4 A separate systematic review reported that the prevalence of any peripartum anxiety disorder was 20.7% (16.7%–25.4%), 30 which is lower than we observed. Previous reports indicated that the incidence of bipolar disorder and psychosis increases postpartum, as we observed, and relapse risk of bipolar disorder increases. 31 , 32 Meta-analyses combined findings across studies using different methods of ascertainment, and rates of mental illness were higher among studies relying on symptom scales than those using diagnostic instruments or administrative data.
Multiple previous studies have demonstrated an increased incidence of mental illness among nonpregnant women with MS. 2 Previously, we reported that people with MS from 4 Canadian provinces had a 71% increased rate of depression, 42% increased rate of anxiety, 99% increased rate of bipolar disorder, and 74% increased rate of schizophrenia compared with age-matched and sex-matched controls; incidence rates were higher in women than in men. 3 Yet, few studies have estimated the incidence or prevalence of perinatal mental illness among mothers with MS. A small study in BC, Canada, used administrative data to investigate the incidence of peripartum mood and anxiety disorders among 255 mothers with MS who had a child born between 1994 and 2006. 7 Overall, 25.7% of mothers with MS had peripartum depression, 28% higher than in mothers without MS. These findings are comparable to those in our Ontario cohort. A more recent study in Norway involving 140 mothers diagnosed with MS before pregnancy were administered the Hopkins Symptom Checklist-25 to assess the prevalence of anxiety and depression once in each of the second and third trimesters of pregnancy and at 6 months postpartum. 9 That study found that 15% of women with MS had depression in the third trimester, 2-fold higher than in women without MS. However, the prevalence of depression postpartum did not differ between groups, nor did the prevalence of anxiety at any point in time. 9 In another study, investigators at a tertiary care center in the United States reviewed medical records to identify depression occurring within the last 4 weeks of pregnancy and the 12-month period after delivery. 33 Of 143 births (111 women), 12.6% were affected by depression, consistent with a high prevalence of peripartum depression as found in our Ontario cohort.
We found that mothers with MS and those with epilepsy had a generally higher incidence and prevalence of mental illness than mothers with IBD or diabetes. Previous small or modestly sized studies have reported a high incidence and prevalence of mental illness among mothers with epilepsy during the peripartum period. In the Norwegian Mother and Child Cohort study, 26.7% of women with epilepsy had prevalent depression and 22.4% had anxiety based on symptom questionnaires during either the PN or postpartum period, which were approximately 50% higher than among mothers without epilepsy. 34 A more recent multicenter study in the United States conducted semistructured psychiatric interviews involving 331 mothers with epilepsy to identify lifetime psychiatric disorders and incident major depression. Compared with 102 healthy mothers, the incidence of major depression did not differ between groups; however, symptoms of depression and anxiety were more severe among mothers with epilepsy. 35 A 2011 systematic review and meta-analysis of peripartum mental illness in chronic medical conditions found a 56% increased risk of PN mental illness and a 52% increased risk of postpartum mental illness overall, but focused on depression and anxiety. 6 Consistent with our findings, women with diabetes had 34% (7%–60%) increased odds of peripartum mental illness overall, whereas women with neurologic disorders other than migraine, including MS (1 study) and epilepsy, had 45% (19%–77%) increased odds. 6 Our findings regarding IBD were similar to those previously reported for Ontario, Canada, using the same data source over the period 2002 to 2015. 36
Strikingly, although only 0.54% of mothers with MS had substance use disorders prenatally, 6% had substance use disorders postpartum. A Manitoba study based on administrative data found that 1.77% of women had a substance use disorder before pregnancy. 12 This decreased to 1.28% during pregnancy and then increased again postpartum to 1.94%; this pattern was similar to what we observed. Administrative data would be expected to underestimate substance use disorders, capturing only those that lead to health care utilization. A survey conducted in the United States between 2012 and 2013 found that the rate of alcohol use disorder was 14.5% among nonpregnant women compared with 9.6% among pregnant women, whereas the rate of drug use was 3.8% in nonpregnant women and 4.1% among pregnant women. 37 These estimates are worrisome because maternal substance use during pregnancy is associated with reduced infant birth weight and head circumference, hypertension, poorer motor performance, and neurologic soft signs such as hyperreflexia, 38 as well as adverse effects on maternal health. 39
Several factors seem to contribute to the increased risk of mental illness among people with MS compared with people without MS including genetic risk of mental illness, 40 exposure to risk factors such as childhood maltreatment, 41 inflammation and immune dysregulation, and brain lesions. 42 Other factors may be relevant during the context of pregnancy, including psychosocial stressors related to pregnancy and chronic disease management; hormonal changes; and the need to adapt to the many physiologic changes that occur during pregnancy 43 including hypercoagulability, insulin resistance, immunotolerance to the fetoplacental unit, and increased plasma volume. The higher incidence of mental illness among mothers with MS and epilepsy than among other mothers may reflect the direct structural, functional, neuroinflammatory, and neurochemical effects of these diseases in the CNS. 44 , 45
These findings have several implications. Untreated maternal depression is associated with substance use, increased risk of future depressive episodes, and suicide. 5 Maternal mental illness also has important effects on the physical and mental health of their child. Mothers with depression breastfeed for shorter periods, and their children receive less preventive care and undergo more emergency department visits. 5 Infants and older children may suffer from behavioral disturbances, developmental delays, and social withdrawal. 5 , 11 Children of affected mothers also develop stress reactivity and altered immune function, 46 which can increase the propensity to chronic illness in childhood or adulthood. 47 , 48
In 2023, the American College of Obstetrics and Gynecology suggested that screening for depression should occur at least twice during pregnancy and postpartum and during pediatric well-infant/child visits, and appropriate treatment initiated. In a systematic review of 50 studies, counseling interventions were associated with a reduced risk of onset of perinatal depression (RR 0.61; 0.47–0.78, number needed to treat: 13.5). 49 For clinicians caring for women with chronic disease, this highlights the importance of being aware of and discussing these issues at the time family planning is taking place, during pregnancy and postpartum.
Strengths of this study include the population-based design involving data from 40% of the Canadian population, supporting generalizability in regions with universal health care systems, consideration of the full spectrum of mental illness, and use of validated case definitions to identify our disease cohorts and mental illness. Limitations should be considered. Even when using validated administrative case definitions, some misclassification of mental illness will occur. For this reason, we used highly sensitive and more specific case definitions, which provide upper and lower bounds for the occurrence of peripartum mental illness. Administrative data only capture conditions for which care is sought; thus, we may have underestimated the burden of mental illness. These data also lack characteristics of the mental illness, including severity, and pharmacologic or nonpharmacologic treatment status, as well as clinical characteristics and health behaviors of the disease cohorts. We did not examine risk factors of mental illness other than the disease cohort, and this should be the subject of future work. Although E-value analysis supports the robustness of most associations observed, residual confounding cannot be excluded. Future work should explore the impact of MS disease activity and severity on peripartum mental illness and whether the intensity of mental health disorders varies among mothers with different disease states.
Our findings suggest that mothers with MS, epilepsy, IBD, and diabetes are at increased risk of peripartum mental illness, across the full spectrum of mental illness; this should be confirmed in clinical data sets. The first postpartum year is a higher risk period for incident mental illness than the PN period. Mothers with MS have higher risks of peripartum depression, anxiety, and psychosis than mothers with other chronic diseases, although these differences were less pronounced for epilepsy, the CNS comparator. Mothers with MS also have a high prevalence of peripartum mental illness. Given the potential adverse consequences for maternal and child health, clinicians caring for mothers with these diseases should be aware of these risks and ensure that recommended screening occurs, followed by appropriate treatment as needed. Greater attention to preventive interventions is also needed.
Acknowledgment
This article used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from Canada Post Corporation and Statistics Canada. Parts of this material are based on data and/or information compiled and provided by CIHI, the MOH, and Statistics Canada. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.
Glossary
Discharge Abstract Database
inflammatory bowel disease
interquartile range
incidence ratio
multiple sclerosis
Ontario Health Insurance Plan
prenatal
positive predictive value
prevalence ratio
rate ratio
Footnotes
Editorial, page e213366
Author Contributions
R.A. Marrie: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. J. Bolton: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. Yushu (Vicki) Ling: drafting/revision of the manuscript for content, including medical writing for content. C. Bernstein: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. K.M. Krysko: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. P. Li: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. K.A. Mckay: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. P. Pequeno: analysis or interpretation of data. N. Razaz: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. D. Rotstein: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. K. Deakin-Harb: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. C.J. Maxwell: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.
Study Funding
This study was funded by MS Canada (EGID 903015). This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care.
Disclosure
R.A. Marrie receives research funding from CIHR, MS Canada, Crohn's and Colitis Canada, the National Multiple Sclerosis Society, CMSC, the Arthritis Society, and the US Department of Defense, and is a coinvestigator on studies receiving funding from Biogen Idec, Roche Canada, and the Pfizer Foundation. J. Bolton receives research funding from MS Canada and the Ruth Hurd Professorship. V. Ling reports no disclosures relevant to the manuscript. C. Bernstein is supported by the Bingham Chair in Gastroenterology. C. Bernstein has served on advisory boards for AbbVie Canada, Amgen Canada, Bristol Myers Squibb Canada, Eli Lilly Canada, Ferring Canada, Innomar Pharmacy, JAMP Pharmaceuticals, Janssen Canada, Pendopharm Canada, Sandoz Canada, Takeda Canada, and Pfizer Canada; received educational grants from AbbVie Canada, Boston Scientific, Eli Lilly Canada, Fresenius Kabi Canada, Ferring Canada, Organon Canada, Pfizer Canada, Takeda Canada, and Janssen Canada; served on speaker's panel for AbbVie Canada, Janssen Canada, Pfizer Canada, and Takeda Canada; and received research funding from AbbVie Canada, Takeda Canada, and Pfizer Canada. K.M. Krysko has received grants from MS Canada and the University of Toronto Division of Neurology; received a contract for a study site from Roche; received speaking or consulting fees from Biogen, EMD Serono, Novartis, and Roche; served as advisory board member for Biogen, EMD Serono, Novartis, and Roche, and served as a scientific advisory committee member for Bristol Myers Squibb. P. Li reports no disclosures relevant to the manuscript. K.A. Mckay receives research funding from the Karolinska Institutet and StratNeuro and has received speaker honoraria from Biogen Inc. and Sanofi. P. Pequeno reports no disclosures relevant to the manuscript. N. Razaz receives funding from the Swedish Research Council. D. Rotstein receives research funding from MS Canada, the National MS Society, the Consortium of Multiple Sclerosis Centers, the University of Toronto Division of Neurology, and Roche Canada; and has received consulting or speaker fees from Alexion, Amgen, Biogen, EMD Serono, Novartis, Roche, Sanofi Aventis, and Touch International Medical Education. K. Deakin-Harb reports no disclosures relevant to the manuscript. C. J. Maxwell receives funding from CIHR, MS Canada, National Multiple Sclerosis Society, the Public Health Agency of Canada, and the Alzheimer Society of Canada. Go to Neurology.org/N for full disclosures.
References
Associated Data
Data Availability Statement
The data set from this study is held securely in coded form at ICES. Legal data-sharing agreements prohibit ICES from making the data set publicly available. However, access may be granted to those who meet prespecified criteria for confidential access, available at ices.on.ca/DAS (email das@ices.on.ca ). The full data set creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are, therefore, either inaccessible or may require modification.
Multiple sclerosis (MS) is a chronic immune-mediated inflammatory disease of the CNS, affecting nearly 3 million individuals worldwide, of whom three-quarters are women. 1 Presentation typically occurs during a period when many individuals are making reproductive decisions. Women with MS have a markedly increased incidence and prevalence of depression, anxiety disorders, and bipolar disorder compared with women without MS. 2 , 3 Peripartum mood disorders constitute the most frequent form of maternal morbidity in the general population 4 and pose health risks for those affected and their families 5 ; however, little is known about the risk of peripartum mental illness in women with MS. A 2017 systematic review of the association between maternal chronic medical conditions and peripartum mental illness identified only 16 studies conducted from 1988 to 2017 6 ; only 1 study included women with MS (n = 255) and focused solely on combined mood and anxiety disorders. 7 A recent scoping review regarding women's health in MS 8 identified only 1 other small study (in Norway) that has directly addressed peripartum mental illness in MS. 9 Among 140 women diagnosed with MS before pregnancy, 15% had depression in their third trimester (based on a screening tool), two-fold more than women without MS; the prevalence of anxiety did not differ. 9 The rising prevalence of MS 1 and number of persons with MS having children 10 and potential adverse effects of peripartum mental illness on the parent and child 5 , 11 make it critical to understand the incidence and prevalence of a broad spectrum of peripartum mental illness in this population.
We compared the incidence and prevalence of peripartum mental illness in women with MS, epilepsy, inflammatory bowel disease (IBD), and diabetes and women without any of these conditions. We chose these conditions because they allow for comparisons on nondemographic factors that may affect the risk of peripartum mental illness including organ system (CNS [MS, epilepsy] vs non-CNS [IBD, diabetes]), pathophysiology (immune-mediated [MS, IBD] vs not [epilepsy, diabetes]), and disease course (relapsing remitting with underlying progression [MS and IBD] vs progressive [diabetes]). We hypothesized that incidence and prevalence of peripartum mental illness among women with MS would exceed those among women in all comparator cohorts except epilepsy, another CNS disorder, and that the incidence of mental illness would be higher in the first postpartum year than in the prenatal (PN) period. 12
Methods
Setting
We conducted this retrospective cohort study in the province of Ontario, which is located in east-central Canada and is home to nearly 40% of the Canadian population (>14 million residents). 13 Ontario has a universal, publicly funded health system with longitudinal, prospectively collected, high-quality records of health service use. 14
Data Sources
Administrative databases capture health service use and are held at ICES (formerly known as the Institute for Clinical and Evaluative Sciences). The Registered Persons Database (RPDB) provided a unique identifier, birth date, sex, region of residence (postal code), and health insurance coverage dates. Postal codes can be linked to census data to determine household income of the individual's enumeration area. The Discharge Abstract Database (DAD) provided hospital admission and separation dates and discharge diagnoses listed using 5-digit International Classification of Diseases (ICD)-9 or 10-Canadian Modification (CA) codes depending on the year. The Ontario Health Insurance Plan (OHIP) database provided dates of physician claims coded using a three-digit ICD-9 code for 1 physician-assigned diagnosis. Physicians paid through alternate funding sources submit shadow billings. 15 The Ontario Mental Health Reporting System captures admissions (discharge diagnoses) to designated adult inpatient mental health beds in general/specialty facilities. 16 Prescription claims data were not accessed because they are not population-based and do not provide indications for use.
The MOMBABY data set has captured all hospital births in Ontario since 1988. As of 2002, mothers and their newborns were deterministically linked by the DAD using the medical chart number. 17 This data set provided gestational age at delivery and indicated whether the birth was live/stillbirth and singleton/multiple. All data sets were linked at the individual level using unique encoded identifiers and analyzed at ICES.
Standard Protocol Approvals, Registrations, and Patient Consents
We obtained ethics approval from the University of Manitoba Health Research Ethics Board. Data access (through ICES) was authorized under section 45 of ON's Personal Health Information Protection Act.
Study Populations
We identified separate chronic disease cohorts (MS, epilepsy, IBD, and diabetes) using validated administrative case definitions 18 – 21 (eTable 1) using all available hospital and OHIP claims for the period April 01, 1988, and December 31, 2020, and excluded anyone missing date of birth or sex. For each chronic disease algorithm, we designated the earliest ICD-9 or 10-CA code identified as the diagnosis date. If women had more than one chronic disease, they were assigned to the disease with the earliest diagnosis date. We linked these cohorts to the MOMBABY data set to identify mothers with a delivery date during the period of interest. Mothers in the MOMBABY data set with a delivery date in the period of interest who did not have MS, epilepsy, IBD, or diabetes were designated as comparators. For each mother, we selected the first pregnancy in the study period. We estimated the conception date by subtracting clinically assessed gestational age, based on a combination of ultrasound and last menstrual period, recorded at the delivery hospitalization from the newborn's birthdate; <0.3% of all records are missing gestational age because this is a mandatory field. 22 The index date was 365 days before the estimated date of conception for the first pregnancy observed ( Figure 1 for design). We further excluded any woman (1) who had an index date before January 1, 2000, or after March 31, 2017; (2) who did not have a valid registration number; (3) whose newborn did not have a valid registration number; (4) who was not a resident in the province and eligible for provincial health insurance for ≥12 months before the index date to allow complete one-year lookback period; (5) with missing sex; (6) younger than 15 years or older than 49 years at conception date; (7) who died before the delivery date; or (7) who had a birth at gestational age <22 weeks (eFigure 1).
Study Design and Periods
Study Period
The earliest index date was January 1, 2000, and the latest was March 31, 2017. This allowed a 1-year lookback period to define covariates (income, rural residence, annual physician visits) at the index date and a 3-year (1,097 days) follow-up period after the delivery date. For each woman, we defined several intervals: the year before conception (prepregnancy), pregnancy (PN period), first year after delivery (postnatal/postpartum), and second year and third year after delivery ( Figure 1 ). Although previous studies of peripartum mental illness in the general population have reported prevalence and, to a lesser extent, incidence estimates, 12 , 23 few have reported which part of the peripartum interval carries the greatest risk. Therefore, we also divided the PN period into trimesters where trimester 1 was from conception date to 13 weeks 6 days, trimester 2 was from 14 weeks to 26 weeks 6 days, and trimester 3 was from 27 weeks to delivery.
Peripartum Mental Illness
We applied validated case definitions for identifying depression, anxiety, bipolar disorder, psychosis, substance use disorder, suicide attempt, and any mental illness (encompassing any of the listed mental illnesses or mental health service use) (eTable 2). The any mental illness definition has a positive predictive value (PPV) of 84.9% and negative predictive value of 96% (eTable 3). This sensitive definition relies on a single diagnosis or mental illness service code to identify mental illness, consistent with the approach of the Canadian Chronic Disease Surveillance System. 24 We used a similar approach for the other mental illnesses; most case definitions exceeded the recommended PPV of 70% (eTable 3). 25 , 26 We also used a specific definition that required multiple diagnosis codes over a specified period (eTable 2).
If a mother met a case definition, we assumed once affected, always affected, in keeping with relapsing-remitting nature of these conditions. To identify mental illness as incident for the PN period, there had to be no diagnosis codes for the mental illness of interest for 2 years before the conception date. For incident mental illness in the postnatal period, there had to be no diagnosis codes for the mental illness of interest for 2 years before the conception date and during the PN period. The 2-year window was based on our previous observation that in immune-mediated inflammatory disease cohorts, among those who met the case definition for depression, the median [interquartile range (IQR)] time between depression claims was 24.0 (36.0–121.0) days; the 99th percentile of the distribution for time between claims was 497 days. 27
Characteristics
Sociodemographic characteristics assessed at index date included income quintile based on postal code and urban vs rural residence based on postal code, where census metropolitan areas with populations ≥100,000 were designated as urban. Maternal age was assessed at conception. Information about ethnicity was not available in the data sets. Pregnancy characteristics included delivery year, parity (any pregnancies before the study period), multiple births, and maternal comorbidity. Comorbidity was assessed using the Obstetric Comorbidity Index, 28 a weighted index. We modified the index to exclude gestational diabetes, to avoid misclassification in the diabetes cohort, and alcohol and drug abuse, because substance use was an outcome. We summarized health care use as the annual number of physician visits at the index and conception dates. We also assessed access to PN care by (1) use of PN ultrasound (present/absent) and (2) number of PN visits for obstetrics reasons, whether to primary care providers or obstetricians.
Analysis
We summarized patient characteristics using means (SDs), median (IQR), and frequency (percentage). We compared characteristics between groups using χ 2 tests, Kruskal-Wallis tests, and analysis of variance (ANOVA). Owing to the large sample size, which may lead to the identification of statistically significant differences between groups that are not clinically meaningful, we show standardized differences rather than p values. Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large. 29
We estimated the crude and age-standardized incidence proportion and prevalence of mental illness during the PN period overall and for each trimester and in the first, second, and third postpartum years. Estimates were standardized to the 2015 Canadian population using the direct method. We compared the age-standardized estimates between cohorts using incidence ratios and prevalence ratios (PRs) and 95% CIs based on bootstrapping with 500 repetitions.
We used a multivariable Poisson regression model to compare the incidence rates between the PN and postnatal periods and between disease groups, using the natural log of person-time as the model offset and scaling the deviance. Covariates included age at conception (15–24, 25–34, ≥35 [reference]), delivery year (2002–2009 [reference], 2010–2019), income (3 lowest quintiles vs 2 highest because some outcomes were rare), and Obstetric Comorbidity Index (0, 1, 2, ≥3). We similarly modeled prevalence using generalized estimating equations with an exchangeable correlation matrix to account for repeated observations over the study period. We limited these regression analyses to any mental illness, depression, and anxiety, given the small cell sizes for other outcomes. We report rate ratios (RRs) and 95% CI.
Complementary Analysis
Given potential differences in the types of health care providers seen and potentially differential recognition of mental illness, we summarized the type of specialist who submitted the first diagnostic claim for the illness (e.g., depression) as follows: primary care provider, psychiatrist, and other specialists. Individuals who were not diagnosed in outpatient settings were categorized as non-OHIP. E-value analysis was used as a sensitivity analysis to show the strength of association required with the disease group and the outcome for an unmeasured confounder to fully explain the observed association and is reported in footnotes of relevant tables and figures.
Statistical analyses used SAS V9.4 (SAS Institute Inc., Cary, NC).
Data Availability
The data set from this study is held securely in coded form at ICES. Legal data-sharing agreements prohibit ICES from making the data set publicly available. However, access may be granted to those who meet prespecified criteria for confidential access, available at ices.on.ca/DAS (email das@ices.on.ca ). The full data set creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are, therefore, either inaccessible or may require modification.
Results
Cohorts
We included 894,852 mothers with live births, 1,745 with MS, 5,954 with epilepsy, 4,924 with IBD, and 13,002 with diabetes; the remaining 869,227 were comparator mothers ( Table 1 , eFigure 1). Among the disease cohort mothers, only 2.9% had more than one of the health conditions of interest. Among mothers with MS, 14 had IBD, 36 had epilepsy, and 35 had diabetes. At conception, mothers with MS were similar in age to mothers with diabetes (average difference = 0.03 years) and slightly older than those with IBD (0.91 years) but substantially older than mothers with epilepsy (3.79 years) and comparator mothers (2.62 years). Mothers with diabetes had the highest comorbidity index scores.
Table 1 Sociodemographic Characteristics of Participants Characteristic Total MS Epilepsy IBD Diabetes Comparators Standardized difference (comparator group as reference) MS Epilepsy IBD Diabetes No. of women 894,852 1,745 5,954 4,924 13,002 869,227 No. of pregnancies during the study period, n (%) 1 422,341 (47.2) 970 (55.6) 3,013 (50.6) 2,393 (48.6) 7,862 (60.5) 408,103 (47.0) 0.17 0.07 0.03 0.27 2 364,651 (40.7) 661 (37.9) 2,248 (37.8) 2,067 (42.0) 4,237 (32.6) 355,438 (40.9) 0.06 0.06 0.03 0.17 ≥3 86,972 (9.7) 114 (6.5) 693 (11.7) 464 (9.4) 903 (6.9) 105,686 (12.2) 0.20 0.02 0.09 0.18 Any pregnancy before index date, n (%) 53,549 (6.0) 119 (6.8) 415 (7.0) 298 (6.1) 2,117 (16.3) 50,510 (5.8) 0.03 0.04 0 0.34 Multiple births (yes) for first pregnancy, n (%) 17,167 (1.9) 51 (2.9) 126 (2.1) 121 (2.5) 300 (2.3) 16,569 (1.9) 0.07 0.01 0.04 0.03 Multiple births (yes) for any pregnancies, n (%) 26,777 (3.0) 73 (4.2) 187 (3.1) 179 (3.6) 409 (3.1) 25,929 (3.0) 0.06 0.01 0.04 0.01 Delivery year (calendar), n (%) 2002–2005 153,932 (17.2) 297 (17.0) 842 (14.1) 770 (15.6) 2,428 (18.7) 149,595 (17.2) 0 0.08 0.04 0.04 2006–2009 244,475 (27.3) 469 (26.9) 1,496 (25.1) 1,266 (25.7) 3,621 (27.8) 237,623 (27.3) 0.01 0.05 0.04 0.01 2010–2013 226,884 (25.4) 462 (26.5) 1,594 (26.8) 1,295 (26.3) 3,202 (24.6) 220,331 (25.3) 0.03 0.03 0.02 0.02 2014–2019 269,561 (30.1) 517 (29.6) 2022 (34.0) 1,593 (32.4) 3,751 (28.8) 261,678 (30.1) 0.01 0.08 0.05 0.03 Age at conception date (yrs), mean (SD) 28.63 (5.7) 31.21 (4.8) 27.42 (5.9) 30.30 (4.8) 31.24 (5.8) 28.59 (5.7) 0.49 0.21 0.31 0.45 Age group at conception, n (%) 15–19 64,468 (7.2) 11 (0.6) 626 (10.5) 61 (1.2) 387 (3.0) 63,383 (7.3) 0.34 0.12 0.3 0.2 20–24 143,209 (16.0) 126 (7.2) 1,319 (22.2) 467 (9.5) 1,228 (9.4) 140,069 (16.1) 0.28 0.16 0.2 0.2 25–29 285,509 (31.9) 509 (29.2) 1,827 (30.7) 1,634 (33.2) 3,269 (25.1) 278,270 (32.0) 0.06 0.03 0.03 0.15 30–34 265,135 (29.6) 652 (37.4) 1,457 (24.5) 1,824 (37.0) 4,262 (32.8) 256,940 (29.6) 0.16 0.12 0.16 0.07 35–39 112,066 (12.5) 377 (21.6) 598 (10.0) 779 (15.8) 2,930 (22.5) 107,382 (12.4) 0.24 0.08 0.1 0.27 40–49 a 24,465 (2.8) 70 (4.0) 127 (2.1) 159 (3.2) 926 (7.1) 23,183 (2.7) 0.07 0.04 0.20 0.03 Region at index, n (%) Urban 807,467 (90.2) 1,601 (91.7) 5,246 (88.1) 4,377 (88.9) 11,863 (91.2) 784,380 (90.2) 0.05 0.07 0.04 0.03 Rural 87,385 (9.8) 144 (8.3) 708 (11.9) 537 (10.9) 1,149 (8.8) 84,847 (9.8) 0.05 0.07 0.04 0.03 Income quintile at index, n (%) Quintile 1 (lowest) 201,140 (22.5) 288 (16.5) 1,550 (26.0) 760 (15.4) 3,253 (25.0) 195,289 (22.5) 0.15 0.08 0.18 0.06 Quintile 2 189,009 (21.1) 344 (19.7) 1,234 (20.7) 909 (18.5) 2,776 (21.4) 183,746 (21.1) 0.04 0.01 0.07 0.01 Quintile 3 185,826 (20.8) 388 (22.2) 1,225 (20.6) 1,071 (21.8) 2,820 (21.7) 180,322 (20.7) 0.04 0 0.02 0.02 Quintile 4 175,579 (19.6) 380 (21.8) 1,094 (18.4) 1,136 (23.1) 2,407 (18.5) 170,562 (19.6) 0.05 0.03 0.08 0.03 Quintile 5 143,298 (16.0) 297 (17.0) 851 (14.3) 1,048 (21.3) 1,746 (13.4) 139,308 (16.0) 0.1 0.05 0.14 0.07 Physician visits, median (IQR) 15 (12–18) 16 (12–19) 16 (13–20) 16 (13–19) 17 (14–21) 15 (12–18) 0.08 0.15 0.13 0.34 Prenatal visits, median (IQR) 15 (12–18) 16 (12–19) 16 (13–20) 16 (13–19) 17 (14–21) 15 (12–18) 0.08 0.15 0.13 0.34 Any prenatal ultrasound, n (%) 785,651 (87.8) 1,586 (90.9) 5,410 (90.0) 4,512 (91.6) 12,092 (93.0) 762,051 (87.7) 0.1 0.1 0.13 0.18 Obstetric Comorbidity Index score at conception, mean (SD) 0.05 (0.31) 0.05 (0.31) 0.08 (0.38) 0.07 (0.37) 0.12 (0.47) 0.05 (0.31) 0 0.07 0.06 0.17 Obstetric Comorbidity Index score at conception, n (%) a 0 858,243 (95.9) 1,678 (96.2) 5,614 (94.3) 4,663 (94.7) 11,861 (91.2) 834,427 (96.0) 0.01 0.08 0.06 0.2 1 30,493 (3.4) 52 (3.0) 274 (4.6) 205 (4.2) 929 (7.1) 29,033 (3.3) 0.02 0.06 0.04 0.17 ≥2 6,116 (0.6) 15 (0.8) 66 (1.2) 56 (1.2) 212 (1.7) 5,767 (0.6) 0.02 0.06 0.06 0.10 Abbreviations: IBD = inflammatory bowel disease; IQR = interquartile range; MS = multiple sclerosis. Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large. a Upper categories collapsed to avoid small cells of <6 and preserve privacy.
Sociodemographic Characteristics of Participants
Abbreviations: IBD = inflammatory bowel disease; IQR = interquartile range; MS = multiple sclerosis.
Standardized differences of 0.20 are considered to be small effects, of 0.5 are medium, and of 0.8 are large.
Upper categories collapsed to avoid small cells of <6 and preserve privacy.
Incident Mental Illness in the MS Cohort
Because age-standardized findings were similar to those from multivariable regression, we present the regression results throughout. The highest crude incidence proportion was for any mental illness; 8.4% of mothers were affected prenatally and 14.2% in the first postpartum year. Among individual disorders, focusing on the specific (more conservative) definitions, the condition with the highest incidence during the PN period was anxiety (3.8%), followed by depression (0.82%) and substance use (0.54%); estimates were higher for sensitive definitions ( Figure 2 ), and no suicide attempts were recorded. Similarly, the pattern of which conditions had the highest and lowest incidence was generally the same during the first, second, and third postpartum years as during the PN period, with the exception of psychosis, where cell sizes were small, and bipolar disorder, which was more common than substance use postpartum. We observed a general pattern of incident mental illness in which the highest risk period was the first postpartum year for all conditions, followed by the PN period; this persisted after multivariable adjustment ( Figure 2 , eTable 4).
Crude Incidence Proportion of Mental Illness by Pregnancy Period in Mothers With Multiple Sclerosis
*Where 2 definitions were tested, only the sensitive definition is presented here. Empty cells are shown where 6 or fewer events were observed to preserve privacy and confidentiality.
Incident Mental Illness in the MS Cohort vs Comparators
In the comparator cohort, the incidence proportion of any mental illness was 7.06% during the PN period and the incidence of anxiety was 2.28% and of depression was 0.95%. Compared with mothers in the comparator cohort, mothers with MS had a higher incidence of PN mental illness (any), and for all individual conditions except suicide attempt for which the sample was too small to estimate accurately ( Figure 3 ). The highest relative ratios were observed for the psychosis definition, followed by the specific definitions of depression and bipolar disorder ( Figure 3 ); ratios were generally higher for specific than sensitive definitions. Incidence proportions among mothers with MS remained higher than in comparator mothers in the first, second, and third postpartum years except for any mental illness in the third postpartum year ( Figure 4 ).
Association Between Mothers With MS, Epilepsy, IBD, and Diabetes, Compared With Controls and Incident Mental Illness in the Prenatal Period (Adjusted Rate Ratios, 95% CI)* , ^
*Adjusted for age at conception, delivery year, income quintile, and Obstetric Comorbidity Index. Where sensitive and specific definitions were tested, only specific definitions were shown for visual simplicity. No suicide attempts were recorded for multiple sclerosis. ^E-value analysis for the smallest statistically significant association of RR 1.13 (lower confidence limit): 1.51 (1.49). Rate ratios of 1.25 have an E-value of 1.81, rate ratios of 1.50 have an E-value of 2.37, and rate ratios of 2.0 have an E-value of 3.41. IBD = inflammatory bowel disease; MS = multiple sclerosis.
Association Between MS and Incident Mental Illness During Prenatal (PN) and First 3 Postpartum Years (PYs), Compared With Controls (Adjusted Rate Ratios, 95% CI)* , ^
*Adjusted for age at conception, delivery year, income quintile, and Obstetric Comorbidity Index. The strength of association is generally higher for specific than sensitive definitions. Suicide attempt was suppressed because of some cells <6. ^E-value analysis for the smallest statistically significant association of RR 1.26 (lower confidence limit): 1.83 (1.46). Rate ratios of 1.50 have an E-value of 2.37, and rate ratios of 2.0 have an E-value of 3.41. IBD = inflammatory bowel disease; MS = multiple sclerosis; Sens = sensitive; Spec = specific.
Other Disease Cohorts
During the PN period, the crude incidence proportion of mental illness and multivariable-adjusted incidence rates were increased in all other disease cohorts for all mental illnesses evaluated ( Figure 3 for adjusted RRs prenatally, eFigure 2 for adjusted RRs in the first postpartum year), with 2 exceptions, namely the incidence proportion of psychosis and suicide attempts in the IBD cohort.
Comparisons Between MS and Other Disease Cohorts: Incident Mental Illness
During the PN period, the MS cohort had an increased incidence of depression (sensitive definition), anxiety, psychosis (specific definition), and substance use compared with the IBD and diabetes cohorts ( Table 2 , eTable 5). The MS cohort also had an increased incidence of depression (both definitions, although not statistically significant for the specific definition) and psychosis (specific definition) compared with the epilepsy cohort during the PN period. Postpartum, the MS cohort continued to demonstrate a general pattern of higher incidence of depression, anxiety, and psychosis than mothers with IBD or diabetes. Findings were more variable when compared with the epilepsy cohort, being similar (any mental illness), higher (depression specific definition, postpartum year 2, bipolar disorder, postpartum year 3), and lower (bipolar disorder specific definition and substance use, postpartum years 1 and 2; psychosis, postpartum years 1–3).
Table 2 Comparison of Incidence Rates of Mental Illness Between Multiple Sclerosis (MS) and Other Disease Cohorts a , b Disease cohort Comparator cohort Prenatal Postpartum y 1 Postpartum y 2 Postpartum y 3 Any mental illness Multiple sclerosis Epilepsy 0.93 (0.81, 1.08) 1.05 (0.93, 1.18) 1.17 (1.00, 1.37) 0.97 (0.81, 1.16) Multiple sclerosis Inflammatory bowel disease 1.12 (0.96, 1.30) 1.03 (0.91, 1.16) 1.08 (0.93, 1.27) 0.85 (0.70, 1.02) Multiple sclerosis Diabetes 1.07 (0.93, 1.23) 1.16 (1.04, 1.29) 1.20 (1.04, 1.39) 1.03 (0.86, 1.22) Depression (specific) Multiple sclerosis Epilepsy 1.15 (0.99, 1.32) 0.98 (0.89, 1.08) 1.20 (1.05, 1.38) 1.14 (0.97, 1.33) Multiple sclerosis Inflammatory bowel disease 1.60 (1.36, 1.87) 1.65 (1.48, 1.84) 1.59 (1.37, 1.85) 0.99 (0.85, 1.16) Multiple sclerosis Diabetes 1.14 (0.99, 1.31) 1.42 (1.29, 1.56) 1.10 (0.97, 1.33) 1.53 (1.32, 1.78) Anxiety Multiple sclerosis Epilepsy 1.07 (0.94, 1.21) 1.24 (1.11, 1.38) 1.03 (0.89, 1.19) 1.07 (0.89, 1.29) Multiple sclerosis Inflammatory bowel disease 1.25 (1.09, 1.43) 0.98 (0.88, 1.09) 1.18 (1.01, 1.37) 1.25 (1.03, 1.52) Multiple sclerosis Diabetes 1.25 (1.10, 1.41) 1.20 (1.08, 1.33) 1.11 (0.97, 1.28) 1.35 (1.13, 1.60) Bipolar disorder (specific) Multiple sclerosis Epilepsy 0.82 (0.71, 0.95) 0.63 (0.55, 0.72) 0.43 (0.34, 0.50) 2.21 (1.90, 2.57) Multiple sclerosis Inflammatory bowel disease 1.20 (1.03, 1.41) 1.02 (0.89, 1.18) 0.48 (0.38, 0.61) 2.17 (1.85, 2.55) Multiple sclerosis Diabetes 1.09 (0.94, 1.26) 1.07 (0.94, 1.22) 0.40 (0.32, 0.50) 1.66 (1.45, 1.90) Psychosis (specific) Multiple sclerosis Epilepsy 1.32 (1.14, 1.54) 1.03 (0.92, 1.14) 0 0.33 (0.27, 0.40) Multiple sclerosis Inflammatory bowel disease 5.76 (4.58, 7.25) 2.89 (2.53, 3.30) 0 0.58 (0.47, 0.71) Multiple sclerosis Diabetes 1.47 (1.27, 1.70) 2.08 (1.87, 2.31) 0 1.19 (0.97, 1.45) Substance use Multiple sclerosis Epilepsy 1.08 (0.93, 1.26) 0.56 (0.46, 0.68) 0.73 (0.61, 0.88) 1.14 (0.99, 1.31) Multiple sclerosis Inflammatory bowel disease 1.81 (1.51, 2.16) 0.60 (0.49, 0.74) 0.84 (0.69, 1.03) 1.61 (1.38, 1.89) Multiple sclerosis Diabetes 1.25 (1.08, 1.46) 0.52 (0.43, 0.63) 0.88 (0.74, 1.06) 1.32 (1.15, 1.51) Abbreviation: RR = rate ratio. a Adjusted for age group, socioeconomic status, delivery year, and Obstetric Comorbidity Index. Rate ratio above 1 indicates increased incidence of the mental illness of interest during the peripartum interval of interest (e.g., prenatal) in mothers with MS vs the comparator disease. Comparisons with controls are given in Figure 4 . b E-value analysis for the smallest statistically significant association of RR 1.16 (lower confidence limit): 1.59 (1.24). Estimates of 1.25 have an E-value of 1.81, of 1.50 have an E-value of 2.37, and of 2.0 have an E-value of 3.41.
Comparison of Incidence Rates of Mental Illness Between Multiple Sclerosis (MS) and Other Disease Cohorts a , b
Abbreviation: RR = rate ratio.
Adjusted for age group, socioeconomic status, delivery year, and Obstetric Comorbidity Index.
Rate ratio above 1 indicates increased incidence of the mental illness of interest during the peripartum interval of interest (e.g., prenatal) in mothers with MS vs the comparator disease. Comparisons with controls are given in Figure 4 .
E-value analysis for the smallest statistically significant association of RR 1.16 (lower confidence limit): 1.59 (1.24). Estimates of 1.25 have an E-value of 1.81, of 1.50 have an E-value of 2.37, and of 2.0 have an E-value of 3.41.
Prevalent Mental Illness in the MS Cohort Compared With Other Cohorts
When we examined prevalence estimates, 42% of mothers with MS had any mental illness during the PN period ( Figure 5 for crude estimates). In the first postpartum year, the prevalence of any mental illness was 50.3%; the most prevalent specific conditions were anxiety (25.1%), followed by depression (4.6%). Notably, approximately 1% of mothers had psychosis, and nearly 6% had substance use disorders in postpartum year 1. Compared with mothers in the comparator cohort, the adjusted prevalence of PN mental illness was generally higher among mothers with MS (eFigure 3), consistent with the findings for incidence. The highest PRs were for substance use and bipolar disorder, followed by psychosis. The prevalence of suicide attempt was reduced. The prevalence of mental illness in the first postpartum year was higher in all non-MS disease cohorts compared with the comparator cohort for all mental illnesses studied with 1 exception (eFigure 4). The prevalence of psychosis (sensitive and specific definitions) was higher in all cohorts except in the IBD cohort, in which it was reduced; however, this should be interpreted with caution because of small numbers.
Crude Period Prevalence of Mental Illness by Pregnancy Period in Mothers With Multiple Sclerosis
Where 2 definitions were tested, the sensitive definition is presented for simplicity. Y= year.
Complementary Analysis
Initial mental illness diagnoses were mostly recorded by family physicians and psychiatrists; although these proportions varied by mental illness, distributions were similar across disease cohorts (eFigure 5).
Discussion
In this large population-based cohort study, we investigated the incidence and prevalence of the spectrum of peripartum mental illness in Ontario women with MS, epilepsy (another CNS disease), IBD (another immune-mediated disease), and diabetes relative to comparators without these conditions. The incidence was nonspecifically increased during the PN and postpartum periods among all disease cohorts examined, affecting nearly all specific mental illnesses examined. Mothers with MS had an increased incidence of depression, anxiety, bipolar disorder, psychosis, and substance use than mothers with IBD and diabetes during the PN period. As hypothesized, the incidence of mental illness was higher in the first postpartum year than during the PN period and continued to decline in the second and third postpartum years.
In a 2017 systematic review, the estimated prevalence of PN depression in high-income countries was 9.2% (8.4%–10%) and of postnatal depression was 9.5% (8.9%–10.1%), 4 comparable to the findings in our comparator cohort. The incidence of PN depression was 13%, higher than we observed, and the estimates of the incidence of postnatal depression ranged widely from 5% to 25%. 4 A separate systematic review reported that the prevalence of any peripartum anxiety disorder was 20.7% (16.7%–25.4%), 30 which is lower than we observed. Previous reports indicated that the incidence of bipolar disorder and psychosis increases postpartum, as we observed, and relapse risk of bipolar disorder increases. 31 , 32 Meta-analyses combined findings across studies using different methods of ascertainment, and rates of mental illness were higher among studies relying on symptom scales than those using diagnostic instruments or administrative data.
Multiple previous studies have demonstrated an increased incidence of mental illness among nonpregnant women with MS. 2 Previously, we reported that people with MS from 4 Canadian provinces had a 71% increased rate of depression, 42% increased rate of anxiety, 99% increased rate of bipolar disorder, and 74% increased rate of schizophrenia compared with age-matched and sex-matched controls; incidence rates were higher in women than in men. 3 Yet, few studies have estimated the incidence or prevalence of perinatal mental illness among mothers with MS. A small study in BC, Canada, used administrative data to investigate the incidence of peripartum mood and anxiety disorders among 255 mothers with MS who had a child born between 1994 and 2006. 7 Overall, 25.7% of mothers with MS had peripartum depression, 28% higher than in mothers without MS. These findings are comparable to those in our Ontario cohort. A more recent study in Norway involving 140 mothers diagnosed with MS before pregnancy were administered the Hopkins Symptom Checklist-25 to assess the prevalence of anxiety and depression once in each of the second and third trimesters of pregnancy and at 6 months postpartum. 9 That study found that 15% of women with MS had depression in the third trimester, 2-fold higher than in women without MS. However, the prevalence of depression postpartum did not differ between groups, nor did the prevalence of anxiety at any point in time. 9 In another study, investigators at a tertiary care center in the United States reviewed medical records to identify depression occurring within the last 4 weeks of pregnancy and the 12-month period after delivery. 33 Of 143 births (111 women), 12.6% were affected by depression, consistent with a high prevalence of peripartum depression as found in our Ontario cohort.
We found that mothers with MS and those with epilepsy had a generally higher incidence and prevalence of mental illness than mothers with IBD or diabetes. Previous small or modestly sized studies have reported a high incidence and prevalence of mental illness among mothers with epilepsy during the peripartum period. In the Norwegian Mother and Child Cohort study, 26.7% of women with epilepsy had prevalent depression and 22.4% had anxiety based on symptom questionnaires during either the PN or postpartum period, which were approximately 50% higher than among mothers without epilepsy. 34 A more recent multicenter study in the United States conducted semistructured psychiatric interviews involving 331 mothers with epilepsy to identify lifetime psychiatric disorders and incident major depression. Compared with 102 healthy mothers, the incidence of major depression did not differ between groups; however, symptoms of depression and anxiety were more severe among mothers with epilepsy. 35 A 2011 systematic review and meta-analysis of peripartum mental illness in chronic medical conditions found a 56% increased risk of PN mental illness and a 52% increased risk of postpartum mental illness overall, but focused on depression and anxiety. 6 Consistent with our findings, women with diabetes had 34% (7%–60%) increased odds of peripartum mental illness overall, whereas women with neurologic disorders other than migraine, including MS (1 study) and epilepsy, had 45% (19%–77%) increased odds. 6 Our findings regarding IBD were similar to those previously reported for Ontario, Canada, using the same data source over the period 2002 to 2015. 36
Strikingly, although only 0.54% of mothers with MS had substance use disorders prenatally, 6% had substance use disorders postpartum. A Manitoba study based on administrative data found that 1.77% of women had a substance use disorder before pregnancy. 12 This decreased to 1.28% during pregnancy and then increased again postpartum to 1.94%; this pattern was similar to what we observed. Administrative data would be expected to underestimate substance use disorders, capturing only those that lead to health care utilization. A survey conducted in the United States between 2012 and 2013 found that the rate of alcohol use disorder was 14.5% among nonpregnant women compared with 9.6% among pregnant women, whereas the rate of drug use was 3.8% in nonpregnant women and 4.1% among pregnant women. 37 These estimates are worrisome because maternal substance use during pregnancy is associated with reduced infant birth weight and head circumference, hypertension, poorer motor performance, and neurologic soft signs such as hyperreflexia, 38 as well as adverse effects on maternal health. 39
Several factors seem to contribute to the increased risk of mental illness among people with MS compared with people without MS including genetic risk of mental illness, 40 exposure to risk factors such as childhood maltreatment, 41 inflammation and immune dysregulation, and brain lesions. 42 Other factors may be relevant during the context of pregnancy, including psychosocial stressors related to pregnancy and chronic disease management; hormonal changes; and the need to adapt to the many physiologic changes that occur during pregnancy 43 including hypercoagulability, insulin resistance, immunotolerance to the fetoplacental unit, and increased plasma volume. The higher incidence of mental illness among mothers with MS and epilepsy than among other mothers may reflect the direct structural, functional, neuroinflammatory, and neurochemical effects of these diseases in the CNS. 44 , 45
These findings have several implications. Untreated maternal depression is associated with substance use, increased risk of future depressive episodes, and suicide. 5 Maternal mental illness also has important effects on the physical and mental health of their child. Mothers with depression breastfeed for shorter periods, and their children receive less preventive care and undergo more emergency department visits. 5 Infants and older children may suffer from behavioral disturbances, developmental delays, and social withdrawal. 5 , 11 Children of affected mothers also develop stress reactivity and altered immune function, 46 which can increase the propensity to chronic illness in childhood or adulthood. 47 , 48
In 2023, the American College of Obstetrics and Gynecology suggested that screening for depression should occur at least twice during pregnancy and postpartum and during pediatric well-infant/child visits, and appropriate treatment initiated. In a systematic review of 50 studies, counseling interventions were associated with a reduced risk of onset of perinatal depression (RR 0.61; 0.47–0.78, number needed to treat: 13.5). 49 For clinicians caring for women with chronic disease, this highlights the importance of being aware of and discussing these issues at the time family planning is taking place, during pregnancy and postpartum.
Strengths of this study include the population-based design involving data from 40% of the Canadian population, supporting generalizability in regions with universal health care systems, consideration of the full spectrum of mental illness, and use of validated case definitions to identify our disease cohorts and mental illness. Limitations should be considered. Even when using validated administrative case definitions, some misclassification of mental illness will occur. For this reason, we used highly sensitive and more specific case definitions, which provide upper and lower bounds for the occurrence of peripartum mental illness. Administrative data only capture conditions for which care is sought; thus, we may have underestimated the burden of mental illness. These data also lack characteristics of the mental illness, including severity, and pharmacologic or nonpharmacologic treatment status, as well as clinical characteristics and health behaviors of the disease cohorts. We did not examine risk factors of mental illness other than the disease cohort, and this should be the subject of future work. Although E-value analysis supports the robustness of most associations observed, residual confounding cannot be excluded. Future work should explore the impact of MS disease activity and severity on peripartum mental illness and whether the intensity of mental health disorders varies among mothers with different disease states.
Our findings suggest that mothers with MS, epilepsy, IBD, and diabetes are at increased risk of peripartum mental illness, across the full spectrum of mental illness; this should be confirmed in clinical data sets. The first postpartum year is a higher risk period for incident mental illness than the PN period. Mothers with MS have higher risks of peripartum depression, anxiety, and psychosis than mothers with other chronic diseases, although these differences were less pronounced for epilepsy, the CNS comparator. Mothers with MS also have a high prevalence of peripartum mental illness. Given the potential adverse consequences for maternal and child health, clinicians caring for mothers with these diseases should be aware of these risks and ensure that recommended screening occurs, followed by appropriate treatment as needed. Greater attention to preventive interventions is also needed.
Acknowledgment
This article used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from Canada Post Corporation and Statistics Canada. Parts of this material are based on data and/or information compiled and provided by CIHI, the MOH, and Statistics Canada. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.
Glossary
Discharge Abstract Database
inflammatory bowel disease
interquartile range
incidence ratio
multiple sclerosis
Ontario Health Insurance Plan
prenatal
positive predictive value
prevalence ratio
rate ratio
Footnotes
Editorial, page e213366
Author Contributions
R.A. Marrie: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. J. Bolton: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. Yushu (Vicki) Ling: drafting/revision of the manuscript for content, including medical writing for content. C. Bernstein: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. K.M. Krysko: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. P. Li: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. K.A. Mckay: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. P. Pequeno: analysis or interpretation of data. N. Razaz: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. D. Rotstein: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. K. Deakin-Harb: drafting/revision of the manuscript for content, including medical writing for content; study concept or design. C.J. Maxwell: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.
Study Funding
This study was funded by MS Canada (EGID 903015). This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care.
Disclosure
R.A. Marrie receives research funding from CIHR, MS Canada, Crohn's and Colitis Canada, the National Multiple Sclerosis Society, CMSC, the Arthritis Society, and the US Department of Defense, and is a coinvestigator on studies receiving funding from Biogen Idec, Roche Canada, and the Pfizer Foundation. J. Bolton receives research funding from MS Canada and the Ruth Hurd Professorship. V. Ling reports no disclosures relevant to the manuscript. C. Bernstein is supported by the Bingham Chair in Gastroenterology. C. Bernstein has served on advisory boards for AbbVie Canada, Amgen Canada, Bristol Myers Squibb Canada, Eli Lilly Canada, Ferring Canada, Innomar Pharmacy, JAMP Pharmaceuticals, Janssen Canada, Pendopharm Canada, Sandoz Canada, Takeda Canada, and Pfizer Canada; received educational grants from AbbVie Canada, Boston Scientific, Eli Lilly Canada, Fresenius Kabi Canada, Ferring Canada, Organon Canada, Pfizer Canada, Takeda Canada, and Janssen Canada; served on speaker's panel for AbbVie Canada, Janssen Canada, Pfizer Canada, and Takeda Canada; and received research funding from AbbVie Canada, Takeda Canada, and Pfizer Canada. K.M. Krysko has received grants from MS Canada and the University of Toronto Division of Neurology; received a contract for a study site from Roche; received speaking or consulting fees from Biogen, EMD Serono, Novartis, and Roche; served as advisory board member for Biogen, EMD Serono, Novartis, and Roche, and served as a scientific advisory committee member for Bristol Myers Squibb. P. Li reports no disclosures relevant to the manuscript. K.A. Mckay receives research funding from the Karolinska Institutet and StratNeuro and has received speaker honoraria from Biogen Inc. and Sanofi. P. Pequeno reports no disclosures relevant to the manuscript. N. Razaz receives funding from the Swedish Research Council. D. Rotstein receives research funding from MS Canada, the National MS Society, the Consortium of Multiple Sclerosis Centers, the University of Toronto Division of Neurology, and Roche Canada; and has received consulting or speaker fees from Alexion, Amgen, Biogen, EMD Serono, Novartis, Roche, Sanofi Aventis, and Touch International Medical Education. K. Deakin-Harb reports no disclosures relevant to the manuscript. C. J. Maxwell receives funding from CIHR, MS Canada, National Multiple Sclerosis Society, the Public Health Agency of Canada, and the Alzheimer Society of Canada. Go to Neurology.org/N for full disclosures.
References
Associated Data
Data Availability Statement
The data set from this study is held securely in coded form at ICES. Legal data-sharing agreements prohibit ICES from making the data set publicly available. However, access may be granted to those who meet prespecified criteria for confidential access, available at ices.on.ca/DAS (email das@ices.on.ca ). The full data set creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are, therefore, either inaccessible or may require modification.